~/writing
Eval Harness
2 posts tagged eval-harness. clear filter
September 2026
- 19How stale eval fixtures pass green while the agent has already changedday 14 · LLM Engineering in 30 Days
Our eval score sat flat and green at 94% for three releases while support tickets tripled — the agent hadn't drifted, the test had.
- 11How data beats seniority in an architecture debateday 5 · Career Growth for Engineers in 30 Days
In one architecture debate, twenty years of seniority lost to a dashboard — and once you see why, you'll start hunting for the one artifact on your team that could make you just as hard to ignore.